EPFL and Lausanne University Hospital (CHUV) have joined a €10-million Africa-EU initiative to scale an AI-powered tool that improves tuberculosis (TB) diagnosis across the continent.
The funding – awarded by the EU’s Global Health EDCTP3 Joint Undertaking – will support the development and deployment of ULTR-AI, a portable AI-based lung ultrasound tool that meets and exceeds WHO standards for TB triage accuracy.
Developed by EPFL’s Laboratory for Intelligent Global Health Technologies and CHUV’s Infectious Diseases Department, ULTR-AI uses a smartphone-connected ultrasound device to detect TB in real-time. It eliminates the need for expensive and complex imaging tools like chest X-rays, making it ideal for frontline health workers in low-resource settings.
“AI has to be accessible, acceptable, and real-world ready,” said EPFL’s Dr. Mary-Anne Hartley. CHUV’s Prof. Noémie Boillat-Blanco adds that the tool addresses a key gap: “Ultrasound interpretation typically requires specialized skills, which frontline health workers often lack.”
The five-year project, known as CAD LUS4TB, begins in Benin, Mali, and South Africa, and will include 3,000 patients. It brings together a consortium of ten institutions across Africa and Europe, including Stellenbosch University, Carnegie Mellon Africa, FIND, and the Swiss Tropical and Public Health Institute.
The AI model will be open access and continually improved through shared clinical data. Beyond TB, the tool can help identify conditions like pneumonia or heart disease, expanding its utility in primary care.
“This project is a major step toward affordable, portable diagnostics that save lives and reduce costs by catching TB earlier,” said Dr. Véronique Suttels, who will co-lead the scientific efforts at EPFL.